Ov er v i ew of Slurry phase bubble column Fisc her T ropsc h Syn thesis reactor and relev an t design parameters

Uugan ba y ar Otgon baatar

Nucle ar Scienc e and Engine ering Dep artment, MIT,Cambrid g e,MA

Abstract

Slurry phase bubble column Fisc her T ropsc h (F-T) r e actor with Iron catalyst is c hosen for con v ersion of syngas in to F-T liquids due to go o d heat and mass transfer prop erties. The parameters that affec t the con v ers io n r ate of CO and molec ular mass distribution of final pro du cts are considered. Calculations are p erformed for heat exc hanged in the reactor and total mass flo w. The corr e latio n b et w een design and p erformance parameters are s ho wn graphically . The rationale b ehind design p a rame ter c hoices are explaine d. Distillation and refining pro cesses explained q ualitativ ely .

1. In tro duction

Slurry phase bubble column t yp e of reactor has b een c hosen for F-T syn thes is pro cess. The slurry phase reactor offers an isothermal op erating condition and b etter heat transfer than alternativ e reactor designs [1]. The sc hematic of slurry phase reactor is sho wn in Figure 1. The reactor v essel is 7m in diameter a n d 30m in heigh t. Syngas en ters from the b ottom of v essel through an inlet nozzle and bubbled v ertically in c h urn turbulen t flo w regime to maximize the mass and heat transfer [2]. The reactor i s filled with inert oil in whic h F e catalyst particles are su sp ended. The heat generated in the c hemical reactions is absorb ed b y a large n um b er of v ertical co olan t tub e s of diameter 50mm. Th e reactor op erates at te mp erature of T= 240 ◦ C and pressure of 24.0 bar . Figure 1 also sho ws the bu bble cross section in slurry . The spacing b e t w een the co olan t tub es is im p ortan t factor that affect the p erformance of the reactor and is discussed in Section 3.

2. Fisc her T ropsc h reactio n and the catalyst

F ollo wing reactions tak e place inside FT reactor

C O + 2 H 2 → − ( C H 2 ) − + H 2 O + 170 k J (1)

H 2 O + C O → C O 2 + H 2 (2)

where − ( C H 2 ) − is meth ylene group and it p olymerizes in to larger molecules[1]. Reaction 2 is c a l led the w ater shift reaction and its rate i s con trolled b y the t yp e of catalyst used whic h ma y or ma y not b e activ e to it. Dep ending on the t yp e catalyst used and C O /H 2 ratio of feed syngas stream, the final FT liquid ma y ha v e differen t para ffi n comp osition. If w e assume that final pro duct consists predominan tly of paraffins or h ydro carb ons with form ula C n H 2 n +2 , then the mass fraction of paraffin with carb on n um b er n in final FT liquid is giv en b y

χ n = n (1 − α AS F ) 2 α n − 1 (3)

where α is the c hain gro wth probabilit y that dep ends on the t yp e of catalyst and C O /H 2 ratio of syngas stream[2]. Figure 2 sho ws the mass fraction of differen t molecules as a function of c hain gro wth probabilit y . In decreasing order of a c tiv it y , the most commonly used catalyst are Ru>F e>Ni>Co>Rh[3]. Although Ru is the most activ e in pro ducing high C n um b er molecules, its price consideration forbids its ec onomic utilization for commercial plan ts. Therefore as the second most activ e catalyst, F e ( α = 0 . 9 ) is c hosen as a cataly s t for F-T reactor[1]. The m a jor supplie r s of catalysts are Dycat In ternational, Haldor T opso e, ICI Katalco and United Catalysts.

Pr eprint submitte d to Elsevier Novemb er 13, 2011

Figure 1: Slurry phase bubb l e reactor sc hematic [1]

Figure 2: ASF distribution for c hain gro wthR.Krishna [1]

Figure 3: Effect of catalyst co ncen tration on con v ersion ratio

3. P erformance parameters

The fraction of CO molecules in feed syngas stream con v erted in to larger paraffins through F-T pro cess is called the con v ersion ratio and is a ma jor determinan t of the p erformance of F-T reactors [1]. The con v ersion factor can b e influence d b y concen tration of catalyst in the slurry , te mp erature a n d the H 2 /C O ratio of feed syngas among other factors. In literature, the total flo w rate of syngas through the reactor is c haracterized b y the sup erficial v elo cit y whic h is defined as the v olumetric flo w rate of syngas p er unit cross sectional area of reactor.

V ˙ n ˙ R T

U s = = = 0 . 05 m/s (4)

A P A

where n ˙ denotes the molar flo w rate and P and A represen t the pressure and the cross sectional area of the reactor resp ectiv ely . Giv en the c hosen parameters for our design the calculation for sup erficial v elo cit y giv es U s = 0 . 12 m/s . The effect of the catalyst concen tration on con v ersion ratio is sho wn in Figure 3. Increasing catalyst concen tration increases con v ersion ratio and ε smax = 0 . 4 is considered the limit for feasible commercial op eration of F-T plan t [2]. F or our design, ε s = 0 . 35 is used to maximize con v ersion within reasonable margin from the lim it.

The n um b er of v ertic a l co olan t tub es needed for heat absorption and to k eep the flo w in c h urn turbulen t flo w regime dep ends o n catalyst concen tration and su p erficial v elo cit y . As illustrated in Figure 3, the n um b er of tub es needed is increased with faster fl o w rate and larger catalyst concen tration. The n um b er of co olan t tub es that corresp onds to the sup erficial v elo cit y of this design is 6000 and pitc h is 0.15m.

H 2 /C O ratio affec ts the the pro duct selectivit y via c hange in c hain gro wth probabilit y α and the reaction temp erature. Song e t al (2004) deriv ed a widely accepted mo de l for c hain gr o wth probabilit y [4].

α = 0 . 2 3 + 0 . 63 · [1 − 0 . 0039( T − 533 K )] (5)

H 2 /C O + 1

Figure 4 and Figure 5 sho ws the effect o f v arying feed ration H 2 /C O and reaction temp erature on selectivit y of F-T syn thesis.

Our design emplo ys feed syngas ratio r = H 2 /C O ≈ 2 . 0 whic h is dictated b y the biomass comp osition. The reaction temp erature is c hosen as T= 245 ◦ C. Ho w ev er, it will b e p ossible to c hange reac tion temp erature b y altering the co olan t flo w rate in reactor.

Heat transfer co efficien t is estimated to b e 1450 W /m 2 from Figure 6. T his g i v es for the heat exc hanged with the co olan t

Q ˙ = 6000 ∗ 30 m ∗ π ∗ 0 . 05 m ∗ 1450 W /m 2 = 41 . 0 M W (6)

Figure 4: Effect of sup erficial v e l o cit y , catalyst concen tration on the n um b er of co olan t tub es R. Krishna [2]

Figure 5: The effect of temp erature on selectivit y at H 2 /C O = 2 . 0

Figure 6: The effect of feed ratio r = H 2 /C O on selectivit y at T= 300 ◦ C

Figure 7: Heat transfer co efficien t as a function of catalyst concen tr a ti on

F or co olan t flo w, w e prop ose to use a one lo op system at pressure P=33.5 bar whic h corresp onds to the saturation pressure at T= 245 ◦ C. The co olan t en ters FT reactor as a condensed saturated liquid at T= 245 ◦ C and lea v es the FT reactor as a saturated v ap or. W e can calculate the mass flo w rate required for co olan t flo w as

m ˙ c =

g

Q ˙

— h f

41 . 0 M W

= = 23 . 5 k g /s 1740 k J /k g

As w e are planning to use riv er w ater for condensing, the tem p erature of the co olan t w ater m ust b e within the limit imp osed b y EP A. The maxim um temp erature that o n e can disc harge co olan t in tak e w ater in to en vironmen t is 20F o v er the in tak e w ater as imp osed b y EP A [5]. W e can then find the co olan t w ater flo w ra te from heat balance as

m i n ˙ tak e = C

Q ˙

w Δ T

= 888 k g /s

(8)

4. Impact of gasifier design on FT pro duct selectivit y

Gasifier pro duces syngas stream whic h is then fed in to FT reactor. Differen t gasifier designs will pro duce syngas streams of differen t comp osition and H2/CO ratio. Since H2/CO ration c ritically impacts pro du ct selectivit y , it is necessary to ev aluate ho w differen t gasifier design c hoices affect final; pro duct comp os ition. T w o d esigns of gasifier are prop ose d and this section pro vides analysis of the final pro duct selectivit y for them. Th e probabilit y of c hain gro wth i s giv en b y follo wing relation.

0 . 23

α = + 0 . 63 · [1 − 0 . 0039( T − 533 K )]

H 2 /C O + 1

T=240C=513K is the op erating temp erature of the reactor and for Silv a gas gasification pro cess whic h is an A merican design, the h ydrogen to CO ratio in the syngas stream is H 2 /C O = 22 / 38 . 2 = 0 . 57 ≈ 0 . 5 . Whereas for S w edish FICBC pro cess H 2 /C O = 4 4 . 4 / 22 . 9 = 1 . 94 ≈ 2 . W e can then compare the c hain gro wth p r obabil it y of the t w o differen t gasifier designs at the same temp erature a s

α S il v a = + 0 . 63 · [1 − 0 . 0039( T − 533 K )] = 0 . 84 (10)

H 2 /C O + 1

α F I C B C = + 0 . 63 · [1 − 0 . 0039( T − 533 K )] = 0 . 76 (11)

H 2 /C O + 1

W e can then apply ASF distribution to calculate mass fraction of molecules with carb o n n um b er n

χ S il v a = 0 . 16 2 n ∗ 0 . 84 n − 1 (12)

χ F I C B C = 0 . 24 2 n ∗ 0 . 76 n − 1 (13)

Figure 1 sho ws mass fr action as a function of carb on n um b er for b oth pro ces ses

Naph tha is a term that refers to mixture of h ydro carb on molecules ha ving carb on n um b ers b et w ee n 5- 12. The term distillate co v ers h ydro carb ons with n b e t w een 12-20. W ax refers to carb on n um b er 20 or more h ydro carb ons. Using this definition, w e can tabulate the mass flo w rate for naph tha, distillate and w ax streams. Silv a gas is c hosen as it pro duces more he a vy diesel pro ducts and is a design that is paten ted in the US. This will impact p ositiv ely on feasibilit y of our design.

Figure 8: Silv a gas and FI C BC for pro duct selectivit y

Mass flo w rates	Silv a gas (%)	FICBC(%)
Naph tha (ga s oline)	52.2	73.5
Distillate(diesel)	31.7	22.1
W ax(diesel)	16	4.3

T able 1: Pro duct mass flo w rates for Silv a and FICBC

5. Distillation

Fisc her T ropsc h liquid is distilled to separate the pro duct in to three differen t pro duct streams according to the molecular mass. The liquid fraction up to 180 ◦ C is s en t out as naph tha, from 180 ◦ C to 320 ◦ C as distillate, and the remainder as a hea vy w ax streamThomas G. Kreutz [6]. Distillation is a standard pro c ess in c hemical en g in eering and design of a distiller is relativ ely simple compared to other units in the whole design. Figure 8 sho ws a simple sc hematic of a generic distiller design. T he v essel cons ists of tra ys f or condensing s eparated pro ducts and heat exc hangers. The three streams of separated FT h ydro carb ons are then sen t to refinery unit.

6. Refining

W e emplo y FTL refining design dev elop ed b y Kreutz et al [6]. Sc hematic of th e refining pro cess is sho w n on Figure 9. FTL is s eparated in to streams of naph tha, distillate, and hea vy w ax b y distillation pro cess and fed directly in to refining. The final pro duct will b e bio disel and bio gasoline blend sto c ks along with ligh t bio gas that is left from distillation or pro duced as a b ypro duct of refining.

Naph tha h ydrotreating and hea vy w ax h ydro crac king are similar subpro cesses that con v ert the organic sulfur a n d nitrogen in feed s tr e am in to H 2 S and N H 3 . Figure 10 sho ws th e reactions that tak e place inside a h ydrotreatmen t reactor. Hydropro cessing step is essen tial for reducing the sulfur con ten t of final fuel pro duct

Figure 9: Distillation unit

Figure 10: Sc hematic of FTL refining pro cess

Figure 11: Hydrotreatmen t reactions and en thalpies

and preparing the feed stream to catalytic reforming stage. In a A stream of high pressure (4bar) h ydrogen is needed for the h ydrotr e atmen t stage and total estimate of h ydrogen is bas ed on refinery design b y Kreutz et al and is 0.1kg/sec [6].

Distillate h ydrotreatmen t is iden tical in principle to naph tha h ydrotreating except that the h ydropro­ cessed d istillate go es directly to die sel p o ol without further refining.

Hydroisomerisation of linear alk a n es in to their higher branc hed isomers in creases the o ctane v alue of the fuel and mak es the com bustion c leaner. In the design w e are considering, the output from naph tha h ydrotreatmen t reactor a n d part of h ydro crac king outputs are fed d ir e ctly in to C 5 /C 6 isomerization unit. The heat of reaction, H , for the h ydroisomerization of C5/C6 paraffin is negligibl e, making it preferable from energy util ization p ersp ec tiv e. High pressu r e h ydrogen stream is supplied to the reaction.

Catalytic reforming is another step to increase the isomerization of the pro du ct for greater o ctane rating. Lo w C n um b er gases are pro duced as a b ypro duct and remo v ed from the pro duct stream. Unlik e h ydroi­ somerization units, the isomerization pro cess is carried out without h ydrogen input to the reactor with the appropriate catalyst.

7. Conclusion

Slurry phase bubble colum n reactor is considered. Design parameters that affect the con v ersion ratio and pro duct selectivit y are the t yp e of catalyst, the concen tration of the catalyst, reactor size, reaction temp erature, feed flo w rate and the H 2 /C O in feed syngas. Dep ending on the final pro duct requiremen t, the conditions of the reactor can b e mo dified to alter the comp osition of pro duct F-T liquid. The para m eters of the re a c to r has b een c hosen so as to maximize the pro duction of carb on n um b er 5-20 molecules whic h can b e used for biogasoline and bio disel pro duction.

Reference

[1] C.Maretto R.Krishna. Mo d elling of bubble column slurry reactor for fisc her- tropsc h syn thesis. Catalysis to day , 52:279–289, 1999.

[2] S.T. Sie R. Krishna. Design and scale-up of the fisc her-tropsc h bubble column slurry reactor. F uel Pr o c essing T e chnolo gy , 64:73–105, 2000.

[3] P .L. Spath and D.C. Da yton. Preliminary screening-tec hnical and economic assessmen t of syn thesis gas to fuels and c hemicals with emphasis on the p oten tial for biomass-deriv ed syngas. T ec hnical rep ort, National Renew able Energy Lab oratory , 2003.

[4] A yhan Demirbas Mrinal K Ghose James G Sp eigh t, Mustafa Balat. The Biofuels handb o ok . RSC publishing, 2011.

[5] Clean w ater act np des p ermitting determinations fo r thermal disc harge and co oling w ater in tak e from miran t k endall station in cam bridge, ma.

[6] Guang jian Liu Rob ert H. Williams Thomas G. Kreu tz, Eric D. Larson. Fisc her-tropsc h fuels from coal and biomass. In Pr o c e e dings of the 25th International Pittsbur gh Co al Confer enc e , 2008.

MIT OpenCourseWare http://ocw.mit.edu

22.033 / 22.33 Nuclear Systems Design Project

Fall 2011

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms .